IDKCS011121
Can we change the Hadoop replication factor of our Hadoop Appliance? WrittenFeb 28, 2018

 
Attachments0
  

Product Categories
Hortonworks

 Question
Can we change the Hadoop replication factor of our Hadoop Appliance?

Environment/Conditions/Configuration

HDP and CDH



Answer
By default, a Teradata Hadoop Appliance will have a set Hadoop replication factor of 3. Modifying the default Hadoop replication factor on a Teradata Hadoop Appliance will make the cluster unsupported and treated as a third party distribution.
 

Special Considerations

Additional Information
Here are some reasons why modifying the Hadoop replication factor is not a good idea:
  • Data drives in the Hadoop Appliance are not configured with RAID to manage data availability because Hadoop manages data availability through data replication.
  • Teradata’s drive replacement strategy is based on a replication factor of three (the Hadoop default).  We do not dispatch a CSR to replace drives each time a drive fails.
  • If the customer changes the replication factor, the customer assumes the risk of data loss.
  • Teradata’s drive replacement policies and practices do not change due to a change in replication factor.
  • Increasing the Hadoop replication factor, and not having enough data nodes to support this configuration, could result in "missing block replicas" warnings.